The Planetary Computer for Studying the Planet Earth

1
The Planetary Computer for Studying the Planet
Earth

• Information Science and Technology Colloquium
• NASA Goddard Space Flight Center
• May 21, 2003

Dr. Larry Smarr Director, California Institute
for Telecommunications and Information
Technologies Harry E. Gruber Professor, Dept. of
Computer Science and Engineering Jacobs School of
Engineering, UCSD
2
Abstract
After 15 years of building out the wired Internet
infrastructure around the world, new extensions
of this core infrastructure are rapidly
emerging.  Wireless Internet access from both
unlicensed and licensed radio bands are spreading
quickly, dedicated dark fiber optical networks
are being set up between universities in many
states and countries, the NSF, NIH, NASA, and
foreign science agencies are starting a decade
long construction of a new generation of shared
scientific facilities to accelerate data
intensive science, federated data systems are
being interlinked, computing is decentralizing
into a Grid framework, and high resolution and
virtual reality interfaces are becoming more
common.  These trends will result in a vastly
more powerful global knowledge and collaboration
system than exists today.  I will give examples
of how the new California Institute for
Telecommunications and Information Technology
(www.calit2.net) is creating "Living
Laboratories" for the earth sciences with its
partner, the UCSD Scripps Institution of
Oceanography, to explore these futures with
interdisciplinary teams. The OptIPuter project,
NSF's largest computer science award from last
year and the Santa Margarita Ecological Reserve
sensornets will be used as specific examples,
along with the NSF EarthScope and GEON knowledge
grids for the earth sciences.
3
The Next S-Curves of Internet Growth A Mobile
Internet Powered by a Planetary Grid

• Wireless Access--Anywhere, Anytime
• Broadband Speeds
• Always Best Connected
• Billions of New Wireless Internet End Points
• Information Appliances
• Sensors and Actuators
• Embedded Processors
• Emergence of a Distributed Planetary Computer
• Parallel Lightwaves in Optical Backbone
• Storage of Data Everywhere
• Scalable Distributed Computing Power

4
Californias Institutes for Science and
Innovation A Bold Experiment in Collaborative
Research
California Institute for Bioengineering,
Biotechnology, and Quantitative Biomedical
Research
Center for Information Technology Research in
the Interest of Society
UCD
UCM
UCB
UCSF
California NanoSystems Institute
UCSC
California Institute for Telecommunications and
Information Technology
www.ucop.edu/california-institutes
5
Cal-(IT)2--An Integrated Approach the Future of
the Internet
220 UC San Diego UC Irvine Faculty Working in
Multidisciplinary Teams With Students, Industry,
and the Community
The States 100 M Creates Unique Buildings,
Equipment, and Laboratories
www.calit2.net
6
The EOSDIS is the Largest e-ScienceDistributed
Information Infrastructure
Flight Operations,
Science Data
Distribution,
Processing,
Access,
Data Capture,
Data
Info Mgmt, Data
Interoperability,
Initial Processing,
Transport
Reuse
Archive, Distribution
Backup Archive
Data Acquisition
to DAACs
Research
Users
Tracking
EOS
Data
Spacecraft
Relay Satellite
(TDRS)
Distributed Active Archive Centers
NASA
Integrated
Data Processing Mission Control
Services
NASA Internet
Network (NISN) Mission Services
Education
Users
White Sands
Complex
(WSC)
Instrument Teams
EOS Polar Ground Stations
Interagency
Data
Centers
7
Distributed Cyberinfrastructure Will Underpin
Many e-Science Community Resources
LHC
ATLAS
8
Distributed Data Grid Supporting International
Particle Physics Experiments
Source Harvey Newman, Caltech
9
NIH is Creating a Federated RepositoryBiomedical
Informatics Research Network
NIH Plans to Expand BIRN to Other Organs and
Many Laboratories
Part of the UCSD CRBS Center for Research on
Biological Structure
National Partnership for Advanced Computational
Infrastructure
10
GEONCyberinfrastructure for the Geosciences
NSF ITR Grant 11.25M 2002-2007
www.geongrid.org

• IT SDSC, Penn State, San Diego State University
• Geosciences Arizona State University, Bryn Mawr
  College, Cornell University, Rice University,
  UNAVCO, University of Arizona, University of
  Idaho, University of Missouri, University of
  Texas El Paso, University of Utah, Virginia Tech
• Education and Outreach DLESE, Cornell, UNAVCO
• Agency Partners USGS, Livermore Labs
• Industry Partners ESRI, GeoFusion, IBM

11
The European Data GridA Universal Information
Infrastructure
12
New Technologies to Improve the Information
Infrastructure

• Dedicated Optical Networks
• Distributed Storage Fabric
• Multi-Mega Pixel Images
• Interactive Volume Visualization
• Integration of Aerial and Satellite Imagery
• Realtime Data fusion of Sensor Nets
• Collaboration Systems

13
Why Optical NetworksAre Emerging as the 21st
Century Driver
Performance per Dollar Spent
0
1
2
3
4
5
Number of Years
Scientific American, January 2001
14
Parallelism Has Come to Optical Networking
Parallel Lambdas Will Drive This Decade The Way
Parallel Processors Drove the 1990s
15
A LambdaGrid Will Create an Optical Fabric as
the Backbone of an e-Science Network
Apps Middleware
Clusters and Storage
C O N T R O L P L A N E
Dynamically Allocated Lightpaths
Switch Fabrics
Physical Monitoring
Source Joe Mambretti, NU
16
The Next S-Curves of NetworkingExponential
Technology Growth

Lambda Grids
Experimental Networks
Production/ Mass Market
DWDM
100
Technology Penetration
Internet2 Abilene
Experimental/ Early Adopters
Connections Program
0
Research
Gigabit Testbeds
Time Technology S-Curve
1990s 2000
2010 Networking Technology S-Curves
17
National Light Rail Footprint Layer 1 Topology
John Silvester, Dave Reese, Tom West-CENIC
18
OptIPuter Data-Intensive Application Drivers
BioScience and GeoScience

• NSF Large Information Technology Research
  Proposal
• UCSD and UIC Lead CampusesLarry Smarr PI
• USC, UCI, SDSU, NW Partnering Campuses
• Industrial Partners IBM, Telcordia/SAIC, Chiaro,
  Calient
• 13.5 Million Over Five Years

NIH Biomedical Informatics Research Network
NSF EarthScope
http//ncmir.ucsd.edu/gallery.html
siovizcenter.ucsd.edu/library/gallery/shoot1/index
.shtml
19
Science Drivers for a Radical New
ArchitectureThe OptIPuter

• Neuro Earth Sciences
• Each Data Object is 3D and Gigabytes
• Data in Distributed Federated Repository
• Want to Interactively Analyze and Visualize
• Need Multiple Disciplinary Specialists
• Science Requirements for Dedicated Link
• Computing Requirements ? PC Clusters
• Communications ? Dedicated Lambdas
• Data ? Large Lambda Attached Storage
• Visualization ? Collaborative Volume Algorithms

Goal Punch a Hole Through the Internet Between
Researchers Lab and Remote Data!
20
The OptIPuter Philosophy
Bandwidth is getting cheaper faster than
storage.Storage is getting cheaper faster than
computing. Exponentials are crossing.
A global economy designed to waste transistors,
power, and silicon area -and conserve bandwidth
above all- is breaking apart and reorganizing
itself to waste bandwidth and conserve power,
silicon area, and transistors." George Gilder
Telecosm (2000)
21
The UCSD OptIPuter Deployment
OptIPuter Dedicated Optical Fiber Campus
Experimental Network
To CENIC
Phase I, Fall 02
Phase I, Fall 02
Phase II, 2003
Phase II, 2003
Collocation point
Collocation point
SDSC
SDSC
SDSCAnnex
SDSCAnnex
Preuss
High School
JSOE
Engineering
CRCA
Arts
SOM
Medicine
6th College
UndergradCollege
Chemistry
Phys. Sci -Keck
Node M
Collocation
SIO
Earth Sciences
Funded by NSF OptIPuter Grant and UCSD
Source Phil Papadopoulos, SDSC Greg Hidley,
Cal-(IT)2
22
Chicago Metro Lambda Switching OptIPuter
Laboratory
Source Tom DeFanti, UIC
23
Photonic Data Services OptIPuter
6. Data Intensive Applications (UCI)
5b. Data Services SOAP, DWTP, (UIC/LAC)
5a. Storage (UCSD)
4. Transport TCP, UDP, SABUL, (USC,UIC)
3. IP
2. Photonic Path Serv. ODIN, THOR,... (NW)
1. Physical
Source Robert Grossman, UIC/LAC
24
Research Into Integrated Optical Control Plane
Source Oliver Yu, UIC
25
OptIPuter QUANTA Middleware Architecture
 
Application Toolkits
Globus High-Level Services (Distributed Computing
Resource Management)
QUANTA Middleware
Integrated Signaling
Dynamic Adaptive Control
Communication
Computation
Intelligent Signaling
Monitoring Instrumentation
Globus Core Services
PIN
ODIN
ELECTRA
OPTICA
Optical Transport Infrastructure
Intra-Cluster
Inter-Cluster
Myrinet
Flexible Lambda Switching
Source Jason Leigh, Oliver Yu, UIC
26
The OptIPuter 2003
Experimental Network Wide Array of Vendors
27
iGrid 2002 Was Sustaining 1-3 Gigabits/s
Total Available Bandwidth Between Chicago and
Amsterdam Was 30 Gigabit/s
28
US Geological Survey Earth Resources
Observation Systems (EROS) Data Center

• Landsat 7 Satellite Program
• Images at 30m Resolution
• Each Scene is 115 x 105 miles
• 6000x6000 Pixels
• 3.8 gigabits per Scene
• 250 Scenes per Day
• The Whole Earth
• 58,000 Scenes
• 220 Terabits
• Moving Just 1 of Earth
• on a 1 Gigabit Network an Hour
• GIS Scientists Need to Have LOTS of Gigabits per
  second!
• EROS is Partnering with Cal-(IT)2

Source Jason Leigh, EVL, UI Chicago
29
Todays Aerial Imaging is 10,000 TimesMore
Detailed than Landsat 7 Satellite Images
Source Eric Frost, SDSU
Shane DeGross
SDSU Campus
30 meter pixels
4 centimeter pixels
Laurie Cooper, SDSU
30
Remote Collaborative Visual Analysis of Large
Datasets

• Driven by SensorNets Data
• Real Time Seismic
• Environmental Monitoring
• Emergency Response
• Distributed Corporations
• Linked UCSD and SDSU
• Dedication March 4, 2002

Linking Control Rooms
UCSD
SDSU
Cox, Panoram, SAIC, SGI, IBM, TeraBurst
Networks SD Telecom Council
44 Miles of Cox Fiber
31
Many Groups Are Building Tiled DisplaysFor Large
2 and 3D Image Viewing
PerspecTile Running JuxtaView Jason Leigh, EVL,
UIC
www.llnl.gov/icc/sdd/img/images/pdf/Walls97.pdf
LCD Panels
Video Projectors
Each 3x520 Megapixels Total
32
Current Bleeding Edge--60 Million PixelsDriven
By Commodity PC Cluster!
At 15 Frames/s, The System Can Display 2.7 GB/Sec
Source Philip D. Heermann, DOE ASCI Program
33
Building a Digital San DiegoLinking GIS Data to
High Res Imagery

• JuxtaView
• Digital Montage Viewer For Perspectile LCD Wall
• 6000x3000 Pixel Resolution
• Display Is Powered By
• 16 PCs With Graphics Cards
• 2 Gigabit Networking Per PC
• USGS EROS Center Data
• 133 Urban Areas
• One Foot Resolution
• 100,000 x 100,000 Pixels for 20 sq.mile Urban Area

Source Jason Leigh, EVL, UIC USGS EROS
34
Integrating Real-Time Geo Changes with SAR and
GPS Geodesy

• Western North America Crustal Dynamics Research
  WInSAR Consortium
• Monitor Strain Accumulation And Release Along The
  North American/Pacific Plate Boundary With An
  Emphasis On The San Andreas Fault Zone
• Crustal Dynamics Community Energized By
  Interferograms Of 1992 Landers Earthquake
• Future-Need Near Near-time Access To SAR Data

(Interferometry - Landers Earthquake Massonnet et
al., 1993)
Source David T. Sandwell, SIO, UCSD
35
Current And Future Earth- Observing SAR Missions

• Data acquisition
• rates 150 Mb/s
• Mission lifetime
• 5-10 years
• Acquisition/processing
• time 1/100

36
From Telephone Conference Calls to Access Grid
International Video Meetings
Can We Create Realistic Telepresence Using
Dedicated Optical Networks?
Access Grid Lead-Argonne NSF STARTAP Lead-UICs
Elec. Vis. Lab
37
Planning for Optically Linking Crisis Management
Control Rooms in California
California Office of Emergency Services,
Sacramento, CA
38
ENDfusion End-to-End Networks for Data Fusion
in a National-Scale Urban Emergency Collaboratory

• ENDfusions planned optical and wireless networks
• The width of the rainbows relates to the amount
  of bandwidth managed as lambdas
• Blue lines are conventional networks

Source Maxine Brown, EVL, UIC
39
A High Definition Access Grid as Imagined In
2007 In A HiPerCollab
SuperHD StreamingVideo
100-Megapixel Tiled Display
Augmented Reality
ENDfusion Project
Source Jason Leigh, EVL, UIC
40
Transitioning to the Always-On Mobile Internet
Two Modes of Wireless Wide Area Cellular
Internet Local Access Wi-Fi
Source Ericsson
41
Required Wireless Services Middleware
Applications
Wireless Services Interface
Mobile Code
Location Awareness
Security
Power Control
Data Management
Real-Time Services
J. Pasquale, UCSD, Cal-(IT)2
42
Using Students to Invent the Futureof Widespread
Use of Wireless Devices

• Broadband Internet Connection via Wireless Wi-Fi
• Over 600 Access Points on the Campus
• Year- Long Living Laboratory Experiment 2001-02
• 500 Computer Science Engineering Undergraduates
• 300 Entering UCSD Sixth College StudentsFall
  2002
• Experiments with Geo-Location and Interactive
  Maps

UC San Diego
UC Irvine
Cal-(IT)2 Team Bill Griswold, Gabriele
Wienhausen, UCSD Rajesh Gupta, UCI
43
The Undergraduates are Explorers of the New
World of Mobile Geolocation
Source Bill Griswold, CSE, UCSD
44
Always-On InternetEnables Continuous Data
Entry and Retrieval
Web Portal Customized to User Device
High speed networking
Networked Storage (SAN)
Storage hardware
Source UCSD/Cal(IT)2
45
NSFs ROADnetBringing SensorNets to the Dirt
Roads and the High Seas

• High Bandwidth Wireless Internet
• Linking Sensors for
• Seismology
• Oceanography
• Climate
• Hydrology
• Ecology
• Geodesy
• Real-Time Data Management
• Joint Collaboration Between
• SIO / IGPP
• UCSD
• SDSC / HPWREN
• SDSU
• Cal-(IT)2 Industrial Cost Sharing

R/V Revelle in Lyttleton, NZ
Santa Margarita Ecological Reserve
http//roadnet.ucsd.edu/
46
Environmental SensorNets--Water and Climate
Instrumentsin the Santa Margarita Ecological
Reserve
Source, Dan Cayan, UCSD SIO
47
Distributed Interactive Video ArraysCoronado
Bridge Demonstration May 15, 2002

• UCSD Team Members
• ROADnet Team
• SDSC, HPWREN
• SIO, Seismic Sensors
• Structural Engineering, Bridge Sensors
• CVRR Lab, Video Arrays
• ONR, SPAWAR

Source Mohan Trivedi, UC San Diego
48
Multi-Media Control RoomUCSD Computer Vision and
Robotics Research Lab
http//hpwren.ucsd.edu/news/020524.html
49
Cal-(IT)2 Homeland Security ExperimentsDuring
Super Bowl 2003
UCSD Worked Closely with SD Police (Chief Maheu)

50
(No Transcript)
51
Goal for This DecadePut Laboratories on a Chip
Source Greg McRae, MIT and ANL
52
Can Schools Become a Source for Data Intensive
Science?
US EPA PAMS Sites
California Air Resources Board
Potential for Dramatic Increase in Coverage!
53
How Can We Make Scientific Discovery as Engaging
as Video Games?
Source Mike Bailey, SDSC
Interactive 3D APPLICATIONS
Underground Earth Sciences
Neurosciences
Anatomy
Geography